I recently created a donchian channel system which tests 14 futures markets. This system is based on a proven system shown to me by a friend, and when I run the backtest, there are no trades placed on any instrument. My intention is to recreate the model system and replicate its results exactly via RightEdge.
I've gone through a number of different routes of adjusting the data, provided by CSI, and even when the system does place trades, the results are nowhere near the test I'm modeling.
What I've done is import the data via "Import Bar and Tick Data", and direct it into a watchlist folder.
After this, I go into "Symbol Information" for each symbol I am testing and do the following:
- Adjust asset type to Future
-Asset currency to USD(all markets traded are North American and trade in USD)
-Tick Size = 0(someone from tradersplace forum directed me to do this, as the tick data would be different back in time due to adjustments)
-Contract Size - typed in by hand according to contract size given on CSI Data Platform. (Eg: Soybean Oil is 60000, Canadian Dollar is 100000, Japanese Yen is 12500000, Cotton is 50000)
-Initial Margin = 0 (someone from tradersplace forum directed me to do this, as it would be difficult to know the margin during past time periods and adjust it)
-Expiration Date = irrelevant
- Set decimal places = 5
I personally feel it has something to do with these settings, and depending on the changes made here, especially margin and contract size, the results come out very differently. I am having some trouble differentiating what is correct or incorrect when it comes to these settings, and so far nothing I have done has even come close to replicating the results of the backtest I am modeling.
I appreciate any insight someone might have on how to correct any errors and successfully replicate this model system.